P. Krishnamurthy
YOU?
Author Swipe
View article: Grounding Large Language Models for Robot Task Planning Using Closed‐Loop State Feedback
Grounding Large Language Models for Robot Task Planning Using Closed‐Loop State Feedback Open
Planning algorithms break complex problems into sequential steps for robots. Recent work employs Large Language Models (LLMs) to generate robot policies directly from natural language in simulation and real‐world settings. Models such as G…
View article: ‘ <b>Love and Wisdom in the Cosmos: A Universal Force</b> ’ by Om Somani, Clever Fox Publishing, Chennai & Bangalore. 2025, 109 p. Rs. 299/-
‘ <b>Love and Wisdom in the Cosmos: A Universal Force</b> ’ by Om Somani, Clever Fox Publishing, Chennai & Bangalore. 2025, 109 p. Rs. 299/- Open
View article: Binary Diff Summarization using Large Language Models
Binary Diff Summarization using Large Language Models Open
Security of software supply chains is necessary to ensure that software updates do not contain maliciously injected code or introduce vulnerabilities that may compromise the integrity of critical infrastructure. Verifying the integrity of …
View article: Ransomware 3.0: Self-Composing and LLM-Orchestrated
Ransomware 3.0: Self-Composing and LLM-Orchestrated Open
Using automated reasoning, code synthesis, and contextual decision-making, we introduce a new threat that exploits large language models (LLMs) to autonomously plan, adapt, and execute the ransomware attack lifecycle. Ransomware 3.0 repres…
View article: SaMOSA: Sandbox for Malware Orchestration and Side-Channel Analysis
SaMOSA: Sandbox for Malware Orchestration and Side-Channel Analysis Open
Cyber-attacks on operational technology (OT) and cyber-physical systems (CPS) have increased tremendously in recent years with the proliferation of malware targeting Linux-based embedded devices of OT and CPS systems. Comprehensive malware…
View article: Towards Effective Offensive Security LLM Agents: Hyperparameter Tuning, LLM as a Judge, and a Lightweight CTF Benchmark
Towards Effective Offensive Security LLM Agents: Hyperparameter Tuning, LLM as a Judge, and a Lightweight CTF Benchmark Open
Recent advances in LLM agentic systems have improved the automation of offensive security tasks, particularly for Capture the Flag (CTF) challenges. We systematically investigate the key factors that drive agent success and provide a detai…
View article: Ramendra Mohan Sinha – A Tribute to an Exploration Geologist Par Excellence
Ramendra Mohan Sinha – A Tribute to an Exploration Geologist Par Excellence Open
View article: A Differentiable Distance Metric for Robotics Through Generalized Alternating Projection
A Differentiable Distance Metric for Robotics Through Generalized Alternating Projection Open
In many robotics applications, it is necessary to compute not only the distance between the robot and the environment, but also its derivative - for example, when using control barrier functions. However, since the traditional Euclidean di…
View article: Chain-of-Frames: Advancing Video Understanding in Multimodal LLMs via Frame-Aware Reasoning
Chain-of-Frames: Advancing Video Understanding in Multimodal LLMs via Frame-Aware Reasoning Open
Recent work has shown that eliciting Large Language Models (LLMs) to generate reasoning traces in natural language before answering the user's request can significantly improve their performance across tasks. This approach has been extende…
View article: OSVI-WM: One-Shot Visual Imitation for Unseen Tasks using World-Model-Guided Trajectory Generation
OSVI-WM: One-Shot Visual Imitation for Unseen Tasks using World-Model-Guided Trajectory Generation Open
Visual imitation learning enables robotic agents to acquire skills by observing expert demonstration videos. In the one-shot setting, the agent generates a policy after observing a single expert demonstration without additional fine-tuning…
View article: RAZER: Robust Accelerated Zero-Shot 3D Open-Vocabulary Panoptic Reconstruction with Spatio-Temporal Aggregation
RAZER: Robust Accelerated Zero-Shot 3D Open-Vocabulary Panoptic Reconstruction with Spatio-Temporal Aggregation Open
Mapping and understanding complex 3D environments is fundamental to how autonomous systems perceive and interact with the physical world, requiring both precise geometric reconstruction and rich semantic comprehension. While existing 3D se…
View article: Proactive Hierarchical Control Barrier Function-Based Safety Prioritization in Close Human-Robot Interaction Scenarios
Proactive Hierarchical Control Barrier Function-Based Safety Prioritization in Close Human-Robot Interaction Scenarios Open
In collaborative human-robot environments, the unpredictable and dynamic nature of human motion can lead to situations where collisions become unavoidable. In such cases, it is essential for the robotic system to proactively mitigate poten…
View article: 3D CAVLA: Leveraging Depth and 3D Context to Generalize Vision Language Action Models for Unseen Tasks
3D CAVLA: Leveraging Depth and 3D Context to Generalize Vision Language Action Models for Unseen Tasks Open
Robotic manipulation in 3D requires learning an $N$ degree-of-freedom joint space trajectory of a robot manipulator. Robots must possess semantic and visual perception abilities to transform real-world mappings of their workspace into the …
View article: Enabling Deep Visibility into VxWorks-Based Embedded Controllers in Cyber-Physical Systems for Anomaly Detection
Enabling Deep Visibility into VxWorks-Based Embedded Controllers in Cyber-Physical Systems for Anomaly Detection Open
We propose the DIVER (Defensive Implant for Visibility into Embedded Run-times) framework for real-time deep visibility into embedded control devices in cyber-physical systems (CPSs). DIVER enables run-time detection of anomalies and targe…
View article: Compliant Control of Quadruped Robots for Assistive Load Carrying
Compliant Control of Quadruped Robots for Assistive Load Carrying Open
This paper presents a novel method for assistive load carrying using quadruped robots. The controller uses proprioceptive sensor data to estimate external base wrench, that is used for precise control of the robot's acceleration during pay…
View article: Explainable AI for Large-Scale Predictive Systems: Techniques, Applications, and Future Directions
Explainable AI for Large-Scale Predictive Systems: Techniques, Applications, and Future Directions Open
This article provides a comprehensive examination of Explainable Artificial Intelligence (XAI) techniques and their applications in large-scale predictive systems. The article explores both model-agnostic and model-specific approaches, exa…
View article: D-CIPHER: Dynamic Collaborative Intelligent Multi-Agent System with Planner and Heterogeneous Executors for Offensive Security
D-CIPHER: Dynamic Collaborative Intelligent Multi-Agent System with Planner and Heterogeneous Executors for Offensive Security Open
Large Language Models (LLMs) have been used in cybersecurity such as autonomous security analysis or penetration testing. Capture the Flag (CTF) challenges serve as benchmarks to assess automated task-planning abilities of LLM agents for c…
View article: SHIELD: Secure Host-Independent Extensible Logging for Tamper-Proof Detection and Real-Time Mitigation of Ransomware Threats
SHIELD: Secure Host-Independent Extensible Logging for Tamper-Proof Detection and Real-Time Mitigation of Ransomware Threats Open
Ransomware's escalating sophistication necessitates tamper-resistant, off-host detection solutions that capture deep disk activity beyond the reach of a compromised operating system while overcoming evasion and obfuscation techniques. To a…
View article: Real-Time Multi-Modal Subcomponent-Level Measurements for Trustworthy System Monitoring and Malware Detection
Real-Time Multi-Modal Subcomponent-Level Measurements for Trustworthy System Monitoring and Malware Detection Open
With increasingly sophisticated cyber-adversaries able to access a wider repertoire of mechanisms to implant malware such as ransomware, CPU/GPU keyloggers, and stealthy kernel rootkits, there is an urgent need for techniques to detect and…
View article: Detecting All-to-One Backdoor Attacks in Black-Box DNNs via Differential Robustness to Noise
Detecting All-to-One Backdoor Attacks in Black-Box DNNs via Differential Robustness to Noise Open
The all-to-one (A2O) backdoor attack is one of the major adversarial threats against neural networks. Most existing A2O backdoor defenses operate in a white-box context, necessitating access to the backdoored model’s architecture, hidden l…
View article: Report: Analysis of Justifications for Flue Gas Desulfurization (FGD) Deadline Extension for Major Indian Power Producers
Report: Analysis of Justifications for Flue Gas Desulfurization (FGD) Deadline Extension for Major Indian Power Producers Open
View article: <b>A Comparative Historiographical Analysis of Socio-Ethical Structures in the Indus Valley Civilization and Sangam Age Keeladi</b>
<b>A Comparative Historiographical Analysis of Socio-Ethical Structures in the Indus Valley Civilization and Sangam Age Keeladi</b> Open
View article: Gender Dynamics in Time Use: A Comparative Analysis of India's 2019 and 2024 Surveys (National-Level Trends)
Gender Dynamics in Time Use: A Comparative Analysis of India's 2019 and 2024 Surveys (National-Level Trends) Open
View article: Comprehensive Analysis of Startup Failures: India, China, and Israel
Comprehensive Analysis of Startup Failures: India, China, and Israel Open
View article: Towards Unified Benchmark and Models for Multi-Modal Perceptual Metrics
Towards Unified Benchmark and Models for Multi-Modal Perceptual Metrics Open
Human perception of similarity across uni- and multimodal inputs is highly complex, making it challenging to develop automated metrics that accurately mimic it. General purpose vision-language models, such as CLIP and large multi-modal mod…
View article: Distributed Inverse Dynamics Control for Quadruped Robots using\n Geometric Optimization
Distributed Inverse Dynamics Control for Quadruped Robots using\n Geometric Optimization Open
This paper presents a distributed inverse dynamics controller (DIDC) for\nquadruped robots that addresses the limitations of existing reactive\ncontrollers: simplified dynamical models, the inability to handle exact\nfriction cone constrai…
View article: Out-of-Distribution Detection with Overlap Index
Out-of-Distribution Detection with Overlap Index Open
Out-of-distribution (OOD) detection is crucial for the deployment of machine learning models in the open world. While existing OOD detectors are effective in identifying OOD samples that deviate significantly from in-distribution (ID) data…
View article: RoboPEPP: Vision-Based Robot Pose and Joint Angle Estimation through Embedding Predictive Pre-Training
RoboPEPP: Vision-Based Robot Pose and Joint Angle Estimation through Embedding Predictive Pre-Training Open
Vision-based pose estimation of articulated robots with unknown joint angles has applications in collaborative robotics and human-robot interaction tasks. Current frameworks use neural network encoders to extract image features and downstr…
View article: Data-Efficient System Identification via Lipschitz Neural Networks
Data-Efficient System Identification via Lipschitz Neural Networks Open
Extracting dynamic models from data is of enormous importance in understanding the properties of unknown systems. In this work, we employ Lipschitz neural networks, a class of neural networks with a prescribed upper bound on their Lipschit…
View article: Collision Avoidance for Convex Primitives via Differentiable Optimization Based High-Order Control Barrier Functions
Collision Avoidance for Convex Primitives via Differentiable Optimization Based High-Order Control Barrier Functions Open
Ensuring the safety of dynamical systems is crucial, where collision avoidance is a primary concern. Recently, control barrier functions (CBFs) have emerged as an effective method to integrate safety constraints into control synthesis thro…